Research, The Field Of Tourism Entity Answering System Based On Sms

Natural language database q&a is a kind of very good q&a patterns.allows to natural language manner database query, take the answer returned to the user.SMS is currently the most convenient information service and communication way. how to take the message and the specific field some fixed business database combination.Realization of natural language database based on text messages, very research value quiz. This paper focus on the field entity based on message natural language database model business data to construct, fdomain entity recognition. SMS theme classification, answer retrieval technology such as carried out a series of research and discussion.Mainly made the following respects achievement:(1) For field database definition and characteristics, business inquires describes an entity and entity attributes to field, attribute the category and the attribute values are in the structure of common operation database specific areas building methods, this method will domain specific entity class business database queries structure definition for entities and entity attributes and entity attributes category, entity attribute value, will entity class business database queries of different attribute is transformed into an entity corresponding attribute query the specific attribute value, may realize the different business sector data storage, general definition of structure and field is irrelevant.(2) To SMS domain business inquires, puts forward the fusion feature classification and theme entity recognition of combining the entity business database entity and attribute category access of SMS search analysis method, this method in SMS characteristics and field entity business database structure characteristics, SMS search analysis problem into a recognition of the field entities and SMS query query problem and theme classification based on the conditions of the areas with the airport named entity recognition methods identify the field entity, using query oriented to the dynamic combination essay classification algorithm for SMS theme classification, extraction field entity attributes category, the answer extraction experiments show that the proposed business database entity and attribute category identification method can increase business data query accuracy.(3) For field entity business database characteristics and SMS characteristics, constructed based on Lucene particular areas entity business database retrieval system, this system answer by the structured business Lucene framework for text database into database, forthcoming database each road record as a text block, using Lucene will all text files, and inverted index structure of database of field entities, and entity attributes categories corresponding listed as index word, based on field hownet phrase similarity calculation methods, calculating inquires the entity name, theme classification and database index word the similarity among the biggest, taking the similarity degree index word retrieval extraction business data query answer, answer retrieval results show that this method can improve the answer inquires the accuracy.(4) To Yunnan travel entity data query, for example, design realized based on text messages of yunnan tourist attractions, hotel q&a prototype system.